1
Taiji practice attenuates psychobiological stress reactivity - a randomized controlled
trial in healthy subjects
Running Title: Taiji practice attenuates psychobiological stress reactivity
Marko Nedeljkovic* a, Brigitte Ausfeld-Hafter
a, Konrad Streitberger
b, Roland Seiler
c, Petra
H. Wirtz d
a University of Bern, Institute of Complementary Medicine, Imhoof-Pavillon, Inselspital, CH-
3010 Bern, Switzerland
b Inselspital Bern, University Department of Anesthesiology and Pain Therapy, CH-3010
Bern, Switzerland
c University of Bern, Institute of Sport Science, Alpeneggstrasse 22, CH-3012 Bern,
Switzerland
d University of Bern, Department of Psychology, Biological and Health Psychology,
Alpeneggstrasse 22, CH-3012 Bern, Switzerland
* Address for correspondence and reprint requests:
Marko Nedeljkovic
University of Bern
Institute of Complementary Medicine KIKOM
Imhoof-Pavillon, Inselspital
3010 Bern, Switzerland
Tel.: +41 31 632 9758
Fax: +41 31 632 4262
Email: [email protected]
2
Abstract
Background: Stress reducing effects of Taiji, a mindful and gentle form of body movement,
have been reported in previous studies, but standardized and controlled experimental studies
are scarce. The present study investigates the effect of regular Taiji practice on
psychobiological stress response in healthy men and women.
Methods: 70 participants were randomly assigned to either Taiji classes or a waiting list.
After 3 months, 26 (8 men, 18 women) persons in the Taiji group and 23 (9 men, 14 women)
in the waiting control group underwent a standardized psychosocial stress test combining
public speaking and mental arithmetic in front of an audience. Salivary cortisol and α-
amylase, heart rate, and psychological responses to psychosocial stress were compared
between the study groups. (ClinicalTrials.gov number, NCT01122706.)
Results: Stress induced characteristic changes in all psychological and physiological
measures. Compared to controls, Taiji participants exhibited a significantly lower stress
reactivity of cortisol (p = .028) and heart rate (p = .028), as well as lower α-amylase levels (p
= .049). They reported a lower increase in perceived stressfulness (p = .006) and maintained a
higher level of calmness (p = .019) in response to psychosocial stress.
Conclusion: Our results consistently suggest that practicing Taiji attenuates psychobiological
stress reactivity in healthy subjects. This may underline the role of Taiji as a useful mind-
body practice for stress prevention.
Keywords
Taiji, psychosocial stress, salivary cortisol, salivary α-amylase, heart rate
3
Introduction
The harmful impact of stress on health has been documented repeatedly (Brotman et al., 2007;
Ehlert et al., 2001; Rozanski et al., 1999; Shonkoff et al., 2009). Stress reactivity research
suggests that adverse consequences of psychosocial stress on physical and mental health may
relate to stress-induced activation of different stress-responsive physiological systems
(Brotman et al., 2007; Ehlert et al., 2001; Lovallo and Gerin, 2003; Raison and Miller, 2003).
Large-magnitude physiological reactions to acute stressors in particular, often combined with
delayed recovery, could be identified as stress-related risk factors for cardiovascular disease
(Chida and Hamer, 2008; Chida and Steptoe, 2010; Steptoe et al., 2006). More precisely,
stress-induced hyperreactivity of the two main human stress systems, the hypothalamus-
pituitary-adrenal (HPA) axis with its end-product cortisol and the sympathetic nervous system
(SNS) may increase cardiovascular risk, either alone and / or by inducing adverse changes in
intermediate biological risk factors for cardiovascular disease such as coagulation activity,
inflammation, or lipids (Brotman et al., 2007; Chida and Steptoe, 2010; Hamer et al., 2010;
Lovallo and Gerin, 2003; Rosmond and Bjorntorp, 2000; Rozanski et al., 2005; Steptoe et al.,
2007; von Kanel et al., 2001).
These findings underline the importance of investigating stress preventive
interventions and their effects on psychobiological stress reactivity. By now, cognitive
behavioural stress management has been repeatedly shown to markedly decrease
psychological and biological reactivity towards psychosocial stress in healthy subjects (Gaab
et al., 2003; Hammerfald et al., 2006). In contrast, the effectiveness of mind-body
interventions for reducing psychosocial stress reactivity so far has only been examined for
compassion meditation, showing a dose dependent effect on immune and psychological stress
responses (Pace et al., 2009).
Mind-body practices are characterized as methods focusing on the interactions among
the brain, mind, body, and behaviour, with the intent of using the mind to affect physical
4
functioning and promoting health (U.S. National Institutes of Health, 2010). In fact, an
increasing amount of scientific evidence suggests that mind-body practices, such as Taiji,
might contribute to improvements in physical and mental health (Jahnke et al., 2010; Klein
and Adams, 2004; Wang et al., 2010a; Wang et al., 2010b). Taiji - variably spelled Taijiquan,
Tai Chi or Tai Chi Chuan - is a mindful and gentle form of slow body movements with roots
in ancient Chinese martial arts. Because of its integration of numerous physical, cognitive,
and contextual components which potentially have independent as well as synergistic
therapeutic value, Taiji has been described as a complex multi-component mind-body practice
(Wayne and Kaptchuk, 2008).
A few studies suggest that Taiji may have stress reducing effects. Two studies suggest
that practicing Taiji has short- and long-term effects on the basal activity of the HPA-axis. In
his pioneering research work, Jin (1989) found that practicing Taiji for 60 min reduced
cortisol levels after Taiji as compared to before. Similarly, results from a non-controlled pilot
study found reduced salivary cortisol levels in healthy subjects, both immediately and four
weeks after they completed a Taiji beginners course (Esch et al., 2007). Hitherto, studies
exploring the suitability of Taiji as a stress management intervention are scarce. In terms of
stress as measured by psychological measures only, a decrease of self-reported stress was
observed in healthy young adults as well as in elderly subjects with cardiovascular disease
risk factors and persons with HIV disease (Esch et al., 2007; Robins et al., 2006; Taylor-Piliae
et al., 2006). To date, one randomized-controlled study assessed the effect of 60 min of Taiji
practice on the psychobiological recovery of subjects after they were exposed to a non-
validated stressor intended to induce mental and emotional stress by having them watch
stressful movies and perform mental arithmetic under time-pressure and noise (Jin, 1992). A
decrease of salivary cortisol was measured after Taiji as well as after three different
interventions (reading, brisk walking, meditation). However, due to limited saliva sampling
and missing pre-stress baseline measurement, findings from this study remain inconclusive.
5
Taken together, studies measuring self-reported psychometric parameters consistently suggest
that Taiji may serve as an effective stress management intervention technique, but its effects
on physiological reactivity to acute stress remain unclear.
To the best of our knowledge, randomized controlled trials examining the effects of
Taiji on physiological and psychological reactivity to standardized psychosocial stressors
have not yet been reported. We thus set out to investigate the effects of Taiji on
psychobiological reactivity to a standardized and well-validated stressor, the Trier Social
Stress Test (TSST). We repeatedly assessed different measures of independent stress
responsive systems such as self-reported stressfulness, mood, and calmness, as well as the
physiological stress indicators salivary cortisol, salivary α-amylase levels and heart rate. We
hypothesized that practicing Taiji would be associated with lower psychobiological stress
reactivity.
Methods
Participants and design
The ethics committee of the Canton of Bern, Switzerland formally approved the research
protocol. Recruitment was carried out from April 2010 to August 2010 through advertisement
of the study on pin boards and on the websites of the University of Bern and the University
Hospital in Bern.
Through telephone screening, healthy subjects aged from 18 to 50 years and fluent in
German were included if the following exclusion criteria did not apply within the six months
prior to the screening (yes/no): regular or occasional intake of any medication, any self-
reported acute or chronic somatic or mental disorders, smoking more than five cigarettes/day,
consumption of more than two alcoholic drinks/day, consuming any kind of addictive
substances, any previous participation in stress research projects (in order to ensure that
6
subjects included were naïve to the TSST protocol), more than one week of predictable
absence during the intervention period, any previous practical experience with Taiji exercises.
Women who were using hormonal contraceptives, were pregnant or planning to become
pregnant during the study were also excluded. The included subjects received complete
written and oral descriptions of the study. Informed written consent was obtained before
participating. After baseline examination was completed, the participants were randomly
assigned to either the Taiji group or the waiting control group. The allocation ratio was 1:1.
Allocation concealment was achieved by using sequentially numbered, opaque and sealed
envelopes. An independent research assistant generated the random allocation sequence by
sealing, mixing and subsequently numbering 80 opaque envelopes. They were opened
individually by the primary investigator (MN) for each eligible subject who had agreed to
participate in the study and completed baseline examination. TSST examination was
completed only on subjects with compliance to start and test instructions. The participant
inclusion process is depicted in Figure 1.
Taiji intervention
The Taiji course being offered to the intervention group started in September 2010 and lasted
for 12 weeks. The training sessions of 60 minutes took place twice a week. Taiji classes
differed in composition (participants chose 2 of 6 potential training time points per week) and
size (5 to 15 participants per session). Participants who missed a class were asked to attend a
make-up class. The intervention group was encouraged to practice Taiji at home in addition to
the classes. The average number of home practice sessions was assessed retrospectively using
a brief self-report questionnaire at the end of the course. Participants’ class attendance was
journalized by the Taiji teacher. All classes were held by the same Taiji teacher. He was
trained in China as well as in Europe, has 10 years of Taiji experience, and is a certified Taiji
7
teacher as awarded by the Swiss Society for Qigong and Taijiquan (Schweizerische
Gesellschaft für Qigong und Taijiquan – SGQT).
In the Taiji course, participants were taught the first 18 sequences of the 37 Chen
Man-Ch’ing Yang-Style Taiji short form. An adaptation of five simplified Taiji movements
from this form has been previously used in Taiji trials on patients with chronic heart failure
(Yeh et al., 2004; Yeh et al., 2011). As our study participants were all healthy, we decided not
to simplify the form but to teach the first 18 movements consecutively, as recommended by
Robinson (2006). The main reasons for choosing this form are the following: (1) inclusion of
the basic Taiji principles such as extension, relaxation and alignment of the body, as well as
holistic and mindful body movements (Wolf et al., 1997), (2) feasibility given the moderate
teaching and practicing time of two hours per week for three months, (3) enhanced
embodiment of basic Taiji principles thanks to frequent repetitions enabled by the shortness
of the 18 sequences. Moreover, the Cheng Man-Ch’ing form is widely taught in Switzerland
and subjects interested in learning the remaining part of this form after the study would easily
find a suitable Taiji-school. Each Taiji session began with warm-up exercises (15 min)
followed by practicing Taiji movements and reviewing the underlying principles (35 min) and
concluded with Taiji related breathing and relaxation exercises (10 min).
Prior to group allocation participants of both study groups were requested not to take
part in any new physical exercise or mind-body program during their study participation. All
participants agreed with this request. After the termination of the study, an equivalent Taiji
training was offered to all subjects participating in the control group.
Assessment of potentially confounding variables
Potential Taiji-related confounding variable. We assessed participants’ previous regular
practical experience (in months) with self-applicable mind-body practices (i.e. meditation,
Feldenkrais, Alexander Technique, Qigong, Yoga, Pilates, guided imagery, deep breathing
8
exercises, progressive muscle relaxation and Reiki) at baseline to rule out a non-Taiji related
influence of prior mind-body practice experience on the parameters under study.
Potential confounders of physiological stress reactivity. We controlled for age (Kudielka et
al., 2004), as well as for menstrual cycle phase (luteal vs. follicular phase, see below) and
gender as salivary cortisol reactivity in hormonal contraceptive-free female subjects is blunted
during the follicular phase and differs from cortisol reactivity in male subjects (Kirschbaum et
al., 1999). We asked all female participants to fill out a questionnaire assessing duration
(days) and regularity of the menstrual cycle phase (yes/no), as well as the dates of onset of
menstruation before and after the stress test examination. Luteal phase was defined as the
time span of 14 days before onset of menstruation (Lenton et al., 1984). Additionally, we
controlled for the cardiovascular risk factors smoking (number of cigarettes smoked per day)
and body mass index (BMI, kg/m2, see Table 1), as well as for regular physical activity
(average hours per week) during the intervention period (Benson et al., 2009; Rimmele et al.,
2007; Rohleder and Kirschbaum, 2006).
Procedure of the Trier social stress test (TSST) examination
The experimental sessions were conducted during the first 3 weeks after termination of the 12
week Taiji intervention between 1300 h and 1800 h. The timing of the stress test performance
was balanced between males and females and between participants in the two study groups.
Participants were told to refrain from eating and drinking anything but water for 2 h and from
intense physical activity, caffeine, nicotine, and alcohol during the 24 h before the
experiment. Participants’ compliance to preparatory instructions and absence of the exclusion
criteria was verified; non-compliant participants were excluded from the TSST. Next, the
ECG recording equipment was attached and the recording was started. We used the Trier
Social Stress Test (TSST) combining a 10 min preparation phase followed by a 5 min mock
job interview, and a 5 min mental arithmetic exercise (Kirschbaum et al., 1993). Both tasks
9
were performed two meters in front of two evaluative panel members dressed in white
laboratory coats, and a conspicuous video camera and microphone. The socio-evaluative
character of this performance task was further underlined by presenting the panel members (a
retired male finance manager and a female psychologist) as experts in evaluation of nonverbal
behaviour. The TSST reliably activates HPA-axis and the sympathetic nervous system
(Dickerson and Kemeny, 2004). During recovery, subjects remained seated in a quiet room
for 60 min.
Outcome measures
All outcomes of interest were measured during the TSST-examination sessions. Physiological
as well as psychometric measures were evaluated. Stress reactivity of repeated salivary
cortisol levels (i.e. the interaction group-by-stress) was defined as the main outcome measure.
Secondary measures included repeated α-amylase, heart rate, and different psychometric
assessment tools.
Physiological Measures. Saliva samples (Salivette ®; Sarstedt AG, Sevelen,
Switzerland) were obtained for determination of salivary cortisol (10 min (-20 min) and 1 min
(-10min) prior to the TSST and immediately (+1 min) as well as 10, 20, 30, 45 and 60 min
after stress cessation) and α-amylase levels (-20 min, -10 min, +1 min, +10 min, +20 min, +45
min). Samples were stored at -20 °C until assaying. After thawing, saliva samples were
prepared for biochemical analysis by centrifuging at 3000 rpm for 5 min to produce a clear
supernatant of low viscosity. Estimation of salivary free cortisol was performed using a
chemiluminescence immunoassay with high sensitivity of 0.16 ng/mL (IBL Hamburg,
Germany). Levels of α-amylase were determined following previously described methods
(Rohleder and Nater, 2009). Both salivary cortisol and salivary α-amylase were analyzed in a
commercial laboratory (Dresden LabService GmbH, Dresden, Germany). Inter- and intra-
assay coefficients of variation were both below 8% (cortisol), and 10% (α-amylase),
10
respectively. A single-channel electrocardiogram (ECG, standard lead) was recorded
continuously at 4036 Hz throughout the experimental session using a portable heart rate (HR)
monitoring device (Medikorder MK3, TOM-Medical, Graz, Austria). HR data was aggregated
to 5 min HR segments. The first 5 min HR segment (-10 min) was defined as baseline. HR
segments measured before (-5 min), during (+5 min, +10 min) and after the stress task (+15
min, +20 min) were considered in statistical analyses.
Psychometric measures. Baseline group characteristics included assessment of
perceived stress and depression symptoms. Perceived stress was assessed by the German
version of the Perceived Stress Scale (PSS) (Cohen and Williamson, 1988). This 10-item self-
report questionnaire measures subjects’ evaluation of the stressfulness of the situations
experienced in the past month of their lives. Items in the PSS were designed to assess how
predictable, uncontrollable and overloading participants consider their lives. Good internal
consistency is reported (Cronbach’s α = .78). Depressive affect and negative thought patterns
were measured by using the “Allgemeine Depressionsskala-Kurzform” (ADS-K)
questionnaire (Hautzinger and Bailer, 1993), the German version of the “Center for
Epidemiological Studies Depression Scale” (CES-D) (Radloff, 1977). This questionnaire was
developed for research in the general population and has shown good internal consistency
(Cronbach’s α = .90). We measured psychological TSST stress reactivity at baseline and
immediately after stress cessation: the Multidimensional Mood Questionnaire (MDMQ)
assesses self-reported mood and calmness with good internal consistencies (“mood” -
Cronbach’s α = .75 to .87; “calmness” - Cronbach’s α = .77 to .83) (Steyer et al., 1997).
Psychological evaluation of perceived stressfulness during the TSST examination was
obtained by completion of a visual analogue scale (VAS) ranging from 0 to 10 with 0
indicating no stress experienced at all.
Statistical analysis
11
Data were analysed using SPSS (version 18) statistical software package for Macintosh (IBM
SPSS Statistics. Somers, NY, USA). The calculation of the optimal total sample size has been
conducted using the statistical software G*Power (Buchner et al., 1997). Based on prior
research on cortisol stress responses reporting effect sizes ranging from f 2
= .28 to .35 (Gaab
et al., 2003; Storch et al., 2007), the optimal total sample size of N = 64 was calculated a
priori to detect an expected medium to large effect size of f 2
= .25 with a power ≥ .85 and α =
.05 (effect size conventions: f 2
: .02 = small, .15 = medium, .35 = large; see Cohen, 1988).
Effect size parameters (f) were calculated from partial η2-values and are reported where
appropriate (effect size conventions: f: .10 = small, .25 = medium, .40 = large; see Cohen,
1988). All analyses were two-tailed, with the level of significance set at p < .05 and the level
of borderline significance at p < .10. Unless indicated, all results are presented as mean
±standard error of means (S.E.M.). Prior to statistical analyses all data were tested for normal
distribution and homogeneity of variance using a Kolmogorov-Smirnov and Levene test. As
cortisol levels were skewed we log-transformed (log10) cortisol data and obtained a normal
distribution. Log-transformed cortisol data were used in statistical analyses but for reasons of
clarity untransformed data are depicted in Figure 2a.
Group characteristics were analyzed by 2-analysis for categorical data, and
independent samples t-test for continuous data. Group differences in TSST related baseline
values were also tested by t-tests.
To reveal possible time and condition effects, repeatedly measured physiological and
psychological data were analyzed by using two way ANCOVAs with repeated measurements
with group as the independent factor (Taiji group vs. control group) and cortisol, heart rate, α-
amylase, perceived stressfulness, mood, and calmness data as repeated dependent factors. We
applied Huynh-Feldt correction where appropriate.
To prevent overcontrolling given our sample size (Babyak, 2004), we performed a
two-step procedure for analyses of physiological parameters. In the first step, representing the
12
main analysis for the primary outcome measure cortisol, we calculated repeated cortisol
ANCOVAs while controlling for cortisol baseline levels, prior experience with self-applicable
mind-body practices, age, menstrual cycle phase, and gender as a priori defined covariates.
Significant effects were further tested in a second step, where we additionally controlled for
smoking and BMI, as well as regular physical activity during the intervention period.
Analyses for α-amylase and heart rate were calculated accordingly. In analyses of repeated
psychological data, we controlled for prior mind-body practice experience as a covariate.
Post-hoc testing of significant effects in the main analyses included separate
recalculation of the previously described ANCOVA analyses for each of the repeated time
points.
Results
Of the 112 subjects who underwent a telephone screening, 40 subjects did not fulfil selection
criteria. Reasons for exclusion and drop-out are documented in Figure 1. Of the remaining 74
subjects, 70 successfully underwent baseline examination and were randomly assigned to
either the Taiji group (N = 35) or to the waiting control group (N = 35). TSST examination
was completed by 26 subjects from the Taiji group (mean age 35.77 1.61; 69% female) and
by 23 subjects from the control group (mean age 35.74 1.31; 61% female) as 21 subjects
dropped out before the TSST or did not fulfil inclusion criteria for the TSST (Fig. 1). Since
95% of all dropouts have not attended the TSST examination, an intention-to-treat approach is
not applicable to this study design. Therefore only subjects completing the TSST were
included in data analysis. We had no missing data. No adverse effects of the Taiji training
were observed or reported.
Group characteristics
13
The two study groups did not significantly differ in group characteristic (Table 1) and drop-
out rate (p = .603). Drop-out subjects did not significantly differ from the subjects completing
the study in any group characteristic (p’s > .415), except BMI (21,17 .49 (drop-out group)
vs. 23,49 .51 (final study group); p < .001).
Physiological stress reactivity
At baseline, the study groups did not differ in cortisol, α-amylase, or heart rate. The TSST
induced significant increases in all physiological measures under study (main effects of stress:
p’s < .001). When controlling for confounders considered in the main analyses, a significant
main effect of stress was observed for cortisol (p < .001) and heart rate (p = .027), but not for
α-amylase (p = .91).
Cortisol. The Taiji group showed an attenuated cortisol stress reactivity as compared to the
control group while controlling for the first set of confounders (i.e. physiological baseline
level, age, gender, menstrual cycle phase, and prior mind-body practice experience)
[interaction group-by-stress: F(2.92/122.50) = 3.16, p = .028, partial η2 = .07, f = .27; main
effect of group: F(1/0.79) = 2.99, p = 0.091; Fig. 2a]. Additional controlling for the second set
of confounders (i.e. smoking status, BMI, and physical activity) did not significantly change
results (p = .044, resp. p = .122). Post-hoc tests revealed a trend towards lower cortisol levels
10 min [F(1/0.22) = 3.80, p = .058], 30 min [F(1/0.20) = 2.96, p = .093], 45 min [F(1/0.20) =
3.71, p = .061] and 60 min after stress cessation [F(1/0.20) = 3.80, p = .058] in the Taiji
group, suggesting a lower increase and a faster recovery of salivary cortisol in the
intervention group (see Fig. 2a).
Alpha-amylase. Compared to controls, participants of the Taiji group showed significant
lower α-amylase activity before and after stress while controlling for the first set of
14
confounders [main effect of group: F(1/100795.10) = 4.12, p = .049, partial η2 = .089, f = .31;
Fig. 2b]. No significant group difference was found for α-amylase stress reactivity
[interaction group-by-stress: p = .16]. After additional consideration of the second set of
confounders the main effect of group remained significant (p = .040) and a trend towards
reduced α-amylase stress reactivity in participants of the Taiji group was revealed [interaction
group-by-stress: p = .086]. Post-hoc testing showed significantly lower α-amylase levels 10
min [F(1/40874.38) = 6.63, p = .014], 20 min [F(1/23952.02) = 4.03, p = .051] and 45 min
after stress cessation [F(1/41612.24) = 8.66, p = .005] in the Taiji group as compared to the
control group, indicating a faster recovery (see Fig. 2b).
Heart rate. The main analysis revealed a significantly blunted heart rate stress reactivity in
the Taiji group compared to the control group [interaction group-by-stress: F(2.55/173.76) =
3.34, p = .028, partial η2 = .087, f = .31 ; Fig. 2c]. Furthermore, the participants of the Taiji
group showed a trend towards lower heart rate levels before and after the stress protocol
[main effect of group: F(1/750.01) = 3.15, p = .083; Fig. 2c]. Additional controlling for the
second set of confounders did not significantly change results. Post-hoc tests revealed that
participants of the Taiji group exhibited significantly lower heart rate levels during the first 5
min [F(1/559.40) = 5.93, p = .019], and the second 5 min [F(1/569.63) = 4.33, p = .044] of the
TSST as compared to controls (see Fig. 2c).
Psychological stress reactivity
At baseline subjective measures of perceived stressfulness, mood, and calmness did not differ
between study groups. The TSST significantly increased perceived stressfulness, worsened
mood, and reduced calmness in all study participants (main effects of stress: p’s < .001).
Controlling for prior experience with self-applicable mind-body practices did not significantly
change findings.
15
Compared to the controls, participants in the Taiji group reported significantly less
stressfulness [F(1/16.37) = 8.48, p = .006, partial η2 = .156, f = .43], maintained a higher level
of calmness [F(1/21.79) = 5.87, p = .019, partial η2 = .113, effect size f = .36] and tended
towards a lower decrease of mood [F(1/17.78) = 3.43, p = .070] in reaction to the TSST (see
Table 2).
Discussion
This is the first randomized-controlled study to explore effects of Taiji on measures of
adrenocortical, autonomic, and psychological responses to a standardized and validated
psychosocial stress task in healthy Taiji beginners. We found for the first time markedly
reduced psychobiological stress responses in Taiji practitioners as compared to a non-Taiji
control group, i.e. attenuated cortisol and heart rate stress reactivity, lower α-amylase levels,
as well as lower perceived stressfulness and better maintenance of calmness in response to the
stress task. Baseline values did not differ between groups and stress induction proved to be
successful, as indicated by the expected significant increases in all physiological measures
under study in the total sample.
The present results extend previous research by suggesting an overall stress-buffering
effect of Taiji practice on a broad array of measures representing different stress-responsive
systems with effect sizes ranging from medium to large. Notably, we recruited Taiji beginners
and our Taiji intervention lasted for 12 weeks. As Taiji is thought to improve its beneficial
effects with increasing practice skills over years (Cheng, 1982), it can be speculated that the
observed effects may be even more pronounced in advanced Taiji practitioners.
In contrast to stress management interventions (Gaab et al., 2003; Hammerfald et al.,
2006; Storch et al., 2007), the Taiji course in our study was neither designed nor taught as a
form of stress management. It was conceptualized to convey an embodiment of basic Taiji
16
principles by applying a guided introspective teaching approach. We did not train any specific
coping strategy (e.g. cognitive restructuring) nor did we use role-plays or psychodrama course
elements as often used in cognitive behavioral (Gaab et al., 2003; Hammerfald et al., 2006)
and resource activating stress management programs (Storch et al., 2007). In contrast to our
Taiji course, training elements of such stress management programs may, in addition to their
specific effects on stress appraisal, have more similarities to the TSST situation and therefore
might additionally prepare for the stress test itself. Considering the lacking emphasis on stress
management in the Taiji intervention, the incongruence between the training environment and
the TSST setting, and the focus on developing Taiji related body awareness and body
mechanics, we feel that the stress protective effects of Taiji observed in our study are likely to
result from a mindful embodiment of effortless stability and calmness in motion. This
reasoning is further supported by our finding that, similarly to the Taiji effect, the control
variable “prior experience with mind-body practices” (other than Taiji) was significantly
associated with blunted cortisol stress reactivity, lower α-amylase levels, as well as lower
perceived stressfulness and better maintenance of mood in response to the stress task (p’s <
.040).
Prior research further supports that the observed attenuation of psychobiological stress
reactivity in our Taiji group may relate to mind-body interaction effects. An increased body
awareness induced by regular Taiji practice has been reported in previous studies (Gyllensten
et al., 2010; Uhlig et al., 2010) and is likely to enhance a resource activating embodiment.
Maintaining resource activating embodiment in turn has been shown to reduce cortisol levels
under resting conditions (Carney et al., 2010). Moreover, coping strategies including
embodiment were an integrated part of a resource activating stress management program
found to attenuate the reactivity of the HPA-axis in response to the TSST (Storch et al.,
2007). As participants did not report any Taiji-participation induced increase in social
contacts (data not shown), it seems unlikely that the observed stress-buffering in the Taiji
17
group relates to a training-induced increase in social support. However, it may be speculated
that subjects in the Taiji group, because of their participation in the active study group, might
have expected to be better prepared for their upcoming performance task and thus achieved
greater emotion regulation during the TSST. To clarify the potential contribution of such an
expectancy effect, future research is needed, preferably by including an additional active
control group with an intervention raising similar expectations.
Our study has several strengths. First, we used a well-validated standardized acute
psychosocial stress task (Dickerson and Kemeny, 2004; Kirschbaum et al., 1993). Second, we
used a non-Taiji control group with randomized assignment. Third, we assessed multiple
parameters indicating reactivity of different independent stress responsive systems. Fourth,
baseline characteristics were thoroughly collected and both study groups were homogenous
regarding their demographic and psychometric parameters, indicating a successful
randomization of subjects. Finally, both groups had moderate scores in questionnaires
assessing baseline levels of perceived stress and depressive affect. It therefore seems unlikely
that the reported results are influenced by pre-existing group differences or selection bias
related to increased proneness to stress.
The following limitations need to be considered. First, our results are restricted to a group of
healthy and well-educated young to middle-aged men and women. They cannot be
generalized to other groups with less advantageous health conditions or social backgrounds.
Second, the retrospectively assessed average number of Taiji home practice sessions per
week, the average time spent on sportive activities per week during the intervention period,
and the determination of the menstrual cycle phase were based on self-report. Third, our
results are restricted to Taiji novices. The effects of long-term Taiji practice on
psychobiological stress-reactivity still need to be investigated. Fourth, our psychobiological
assessment approach does not include assessment of further stress-responsive physiological
systems such as the immune, the lipid, or the coagulation system. Also, HPA axis measures
18
other than cortisol such as corticotropin releasing hormone (CRH), or adrenocorticotrophic
hormone (ACTH) still need to be examined. Fifth, because of habituation of cortisol
responses in the majority of people repeating the TSST (Schommer et al., 2003), it was not
possible to assess cortisol stress responses before and after the intervention, nor in the control
group after completion of their Taiji course. Sixth, Our non-significant effects of Taiji on α-
amylase stress reactivity should be interpreted with care as we cannot rule out a type II error.
Future studies, preferably with a higher power, are needed to confirm our non-significant
effects of Taiji on α-amylase stress reactivity as well as the overall stress-reducing effects on
the other measures under study. Finally, due to our restrictive exclusion criteria we had a
drop-out rate of 30%. However, this rate is comparable to a prior TSST study examining mind
body practices (Pace et al., 2009) and does not seem unusual in studies examining Taiji
effects on psychological well-being (Wang et al., 2010a).
In conclusion, our results consistently suggest that practicing Taiji attenuates
psychobiological stress reactivity. This may underline the role of Taiji as a useful mind-body
practice for stress prevention which may contribute to enhance health in the general
population. Clinical implications remain to be elucidated.
Conflict of interest
All authors declare that they have no conflicts of interest.
Role of funding source
Funding for this study was provided by Stiftung für Komplementärmedizin, Gottfried und
Julia Bangerter-Rhyner Stiftung and Parrotia Stiftung (to MN) and by the Swiss National
Foundation Grant PP00P1_128565/1 (to PHW). The funding sources had no further role in
study design; in the collection, analysis and interpretation of data; in the writing of the report;
and in the decision to submit the paper for publication.
19
Acknowledgements
Funding for this study was provided by Stiftung für Komplementärmedizin, Gottfried und
Julia Bangerter-Rhyner Stiftung and Parrotia Stiftung (to MN) and by the Swiss National
Foundation Grant PP00P1_128565/1 (to PHW).
We thank Barbara Schwab, Christina Bürgler, Tina Camenzind, Marina Haller and
Nikola Nedeljkovic for their assistance in conducting the study, and Dr. Ursula Wolf and
Dietrich von Bonin for their technical support and helpful advises.
Contributors
MN was involved in study design, writing of the proposal for the ethics committee, enrolment
of participants, supervising the intervention, data collection, data analysis, and writing of the
manuscript. BA was involved in study design, writing of the proposal for the ethics
committee, and manuscript reviewing; KS and RS were involved in study design, and
manuscript reviewing; PW was involved in study design, data analysis, and writing of the
manuscript. All authors contributed to and have approved the final manuscript.
20
References
Babyak, M.A., 2004. What you see may not be what you get: a brief, nontechnical
introduction to overfitting in regression-type models. Psychosom Med, 66, 411–421.
Benson, S., Arck, P.C., Tan, S., Mann, K., Hahn, S., Janssen, O. E., Schedlowski, M.,
Elsenbruch, S., 2009. Effects of obesity on neuroendocrine, cardiovascular, and
immune cell responses to acute psychosocial stress in premenopausal women.
Psychoneuroendocrinology, 34, 181–189.
Brotman, D.J., Golden, S.H., Wittstein, I.S., 2007. The cardiovascular toll of stress. Lancet,
370, 1089–1100.
Buchner, A., Faul, F., Erdfelder, E., 1997. G*Power: A priori, post-hoc, and compromise
power analyses for the Macintosh (Version 2.1.2). Trier, Germany: University of
Trier.
Carney, D.R., Cuddy, A.J., Yap, A.J., 2010. Power posing: brief nonverbal displays affect
neuroendocrine levels and risk tolerance. Psychol Sci, 21, 1363–1368.
Cheng, M., 1982. Master Cheng's thirteen chapters on T'ai Chi Chuan. Sweet Chi Press, New
York.
Chida, Y., Hamer, M., 2008. Chronic psychosocial factors and acute physiological responses
to laboratory-induced stress in healthy populations: a quantitative review of 30 years
of investigations. Psychol Bull, 134, 829–885.
Chida, Y., Steptoe, A., 2010. Greater cardiovascular responses to laboratory mental stress are
associated with poor subsequent cardiovascular risk status: a meta-analysis of
prospective evidence. Hypertension, 55, 1026–1032.
Cohen, J., 1988. Statistical power analysis for the behavior sciences. Lawrence Earlbaum
Associates, Hillsdale, New Jersey.
21
Cohen, S., Williamson, G., 1988. Perceived stress in a probability sample of the United
States. In: Spacapan, S., Oskamp, S. (Eds.), The social psychology of health:
Claremont Symposium on applied social psychology. Sage Publications, Newsbury
Park, CA, pp. 31–67.
Dickerson, S.S., Kemeny, M.E., 2004. Acute stressors and cortisol responses: a theoretical
integration and synthesis of laboratory research. Psychol Bull, 130, 355–391.
Ehlert, U., Gaab, J., Heinrichs, M., 2001. Psychoneuroendocrinological contributions to the
etiology of depression, posttraumatic stress disorder, and stress-related bodily
disorders: the role of the hypothalamus-pituitary-adrenal axis. Biol Psychol, 57, 141–
152.
Esch, T., Duckstein, J., Welke, J., Braun, V., 2007. Mind/body techniques for physiological
and psychological stress reduction: stress management via Tai Chi training - a pilot
study. Med Sci Monit, 13, CR488–497.
Gaab, J., Blattler, N., Menzi, T., Pabst, B., Stoyer, S., Ehlert, U., 2003. Randomized
controlled evaluation of the effects of cognitive-behavioral stress management on
cortisol responses to acute stress in healthy subjects. Psychoneuroendocrinology, 28,
767–779.
Gyllensten, A.L., Hui-Chan, C.W., Tsang, W.W., 2010. Stability limits, single-leg jump, and
body awareness in older Tai Chi practitioners. Arch Phys Med Rehabil, 91, 215–220.
Hamer, M., O'Donnell, K., Lahiri, A., Steptoe, A., 2010. Salivary cortisol responses to mental
stress are associated with coronary artery calcification in healthy men and women. Eur
Heart J, 31, 424–429.
Hammerfald, K., Eberle, C., Grau, M., Kinsperger, A., Zimmermann, A., Ehlert, U., Gaab, J.,
2006. Persistent effects of cognitive-behavioral stress management on cortisol
responses to acute stress in healthy subjects--a randomized controlled trial.
Psychoneuroendocrinology, 31, 333–339.
22
Hautzinger, M., Bailer, M., 1993. Allgemeine Depressions-Skala [Center for Epidemiological
Studies Depression Scale]. Beltz, Weinheim.
Jahnke, R., Larkey, L., Rogers, C., Etnier, J., Lin, F., 2010. A comprehensive review of health
benefits of qigong and tai chi. Am J Health Promot, 24, e1–e25.
Jin, P., 1989. Changes in heart rate, noradrenaline, cortisol and mood during Tai Chi. J
Psychosom Res, 33, 197–206.
Jin, P., 1992. Efficacy of Tai Chi, brisk walking, meditation, and reading in reducing mental
and emotional stress. J Psychosom Res, 36, 361–370.
Kirschbaum, C., Kudielka, B.M., Gaab, J., Schommer, N.C., Hellhammer, D.H., 1999. Impact
of gender, menstrual cycle phase, and oral contraceptives on the activity of the
hypothalamus-pituitary-adrenal axis. Psychosom Med, 61, 154–162.
Kirschbaum, C., Pirke, K.M., Hellhammer, D.H., 1993. The 'Trier Social Stress Test'--a tool
for investigating psychobiological stress responses in a laboratory setting.
Neuropsychobiology, 28, 76–81.
Klein, P.J., Adams, W.D., 2004. Comprehensive therapeutic benefits of Taiji: a critical
review. Am J Phys Med Rehabil, 83, 735–745.
Kudielka, B.M., Buske-Kirschbaum, A., Hellhammer, D.H., Kirschbaum, C., 2004. HPA axis
responses to laboratory psychosocial stress in healthy elderly adults, younger adults,
and children: impact of age and gender. Psychoneuroendocrinology, 29, 83–98.
Lenton, E.A., Landgren, B.M., Sexton, L., 1984. Normal variation in the length of the luteal
phase of the menstrual cycle: identification of the short luteal phase. Br J Obstet
Gynaecol, 91, 685–689.
Lovallo, W.R., Gerin, W., 2003. Psychophysiological reactivity: mechanisms and pathways to
cardiovascular disease. Psychosom Med, 65, 36–45.
23
Pace, T.W., Negi, L.T., Adame, D.D., Cole, S.P., Sivilli, T.I., Brown, T.D., Issa, M.J., Raison,
C.L., 2009. Effect of compassion meditation on neuroendocrine, innate immune and
behavioral responses to psychosocial stress. Psychoneuroendocrinology, 34, 87–98.
Radloff, L.S., 1977. The CES-D Scale: A self-report depression scale for research in the
general population. Applied Psychological Measurement, 1, 385–401.
Raison, C.L., Miller, A.H., 2003. When not enough is too much: the role of insufficient
glucocorticoid signaling in the pathophysiology of stress-related disorders. Am J
Psychiatry, 160, 1554–1565.
Rimmele, U., Zellweger, B.C., Marti, B., Seiler, R., Mohiyeddini, C., Ehlert, U., Heinrichs,
M., 2007. Trained men show lower cortisol, heart rate and psychological responses to
psychosocial stress compared with untrained men. Psychoneuroendocrinology, 32,
627–635.
Robins, J.L., McCain, N.L., Gray, D.P., Elswick, R.K., Walter, J.M., McDade, E., 2006.
Research on psychoneuroimmunology: Tai Chi as a stress management approach for
individuals with HIV disease. Appl Nurs Res, 19, 2–9.
Robinson, R., 2006. Tai Chi for you - the comprehensive guide to Tai Chi at home for
everybody. Duncan Braid Publishers, London.
Rohleder, N., Kirschbaum, C., 2006. The hypothalamic-pituitary-adrenal (HPA) axis in
habitual smokers. Int J Psychophysiol, 59, 236–243.
Rohleder, N., Nater, U. M., 2009. Determinants of salivary alpha-amylase in humans and
methodological considerations. Psychoneuroendocrinology, 34, 469–485.
Rosmond, R., Bjorntorp, P., 2000. The hypothalamic-pituitary-adrenal axis activity as a
predictor of cardiovascular disease, type 2 diabetes and stroke. J Intern Med, 247,
188–197.
Rozanski, A., Blumenthal, J.A., Davidson, K.W., Saab, P.G., Kubzansky, L., 2005. The
epidemiology, pathophysiology, and management of psychosocial risk factors in
24
cardiac practice: the emerging field of behavioral cardiology. J Am Coll Cardiol, 45,
637–651.
Rozanski, A., Blumenthal, J. A., Kaplan, J., 1999. Impact of psychological factors on the
pathogenesis of cardiovascular disease and implications for therapy. Circulation, 99,
2192–2217.
Schommer, N.C., Hellhammer, D.H., Kirschbaum, C., 2003. Dissociation between reactivity
of the hypothalamus-pituitary-adrenal axis and the sympathetic-adrenal-medullary
system to repeated psychosocial stress. Psychosom Med, 65, 450–460.
Shonkoff, J.P., Boyce, W.T., McEwen, B.S., 2009. Neuroscience, molecular biology, and the
childhood roots of health disparities: building a new framework for health promotion
and disease prevention. JAMA, 301, 2252–2259.
Steptoe, A., Donald, A.E., O'Donnell, K., Marmot, M., Deanfield, J.E., 2006. Delayed blood
pressure recovery after psychological stress is associated with carotid intima-media
thickness: Whitehall psychobiology study. Arterioscler Thromb Vasc Biol, 26, 2547–
2551.
Steptoe, A., Hamer, M., Chida, Y., 2007. The effects of acute psychological stress on
circulating inflammatory factors in humans: a review and meta-analysis. Brain Behav
Immun, 21, 901–912.
Steyer, R., Schwenkmezger, P., Notz, P., Eid, M., 1997. Der Mehrdimensionale
Befindlichkeitsfragebogen (MDBF) [MDMQ - Multidimensional Mood
Questionnaire]. Hogrefe, Göttingen.
Storch, M., Gaab, J., Kuttel, Y., Stussi, A.C., Fend, H., 2007. Psychoneuroendocrine effects
of resource-activating stress management training. Health Psychol, 26, 456–463.
Taylor-Piliae, R.E., Haskell, W.L., Waters, C.M., Froelicher, E.S., 2006. Change in perceived
psychosocial status following a 12-week Tai Chi exercise programme. J Adv Nurs, 54,
313–329.
25
U.S. National Institutes of Health, 2010. http://nccam.nih.gov/health/whatiscam/
Uhlig, T., Fongen, C., Steen, E., Christie, A., Odegard, S., 2010. Exploring Tai Chi in
rheumatoid arthritis: a quantitative and qualitative study. BMC Musculoskelet Disord,
11, 43.
von Kanel, R., Mills, P.J., Fainman, C., Dimsdale, J.E., 2001. Effects of psychological stress
and psychiatric disorders on blood coagulation and fibrinolysis: a biobehavioral
pathway to coronary artery disease? Psychosom Med, 63, 531–544.
Wang, C., Bannuru, R., Ramel, J., Kupelnick, B., Scott, T., Schmid, C.H., 2010a. Tai Chi on
psychological well-being: systematic review and meta-analysis. BMC Complement
Altern Med, 10, 23.
Wang, C., Schmid, C.H., Rones, R., Kalish, R., Yinh, J., Goldenberg, D. L., Lee, Y.,
McAlindon, T., 2010b. A randomized trial of tai chi for fibromyalgia. N Engl J Med,
363, 743–754.
Wayne, P.M., Kaptchuk, T.J., 2008. Challenges inherent to T'ai Chi research: part I--T'ai Chi
as a complex multicomponent intervention. J Altern Complement Med, 14, 95–102.
Wolf, S.L., Coogler, C., Xu, T., 1997. Exploring the basis for Tai Chi Chuan as a therapeutic
exercise approach. Arch Phys Med Rehabil, 78, 886–892.
Yeh, G.Y., McCarthy, E.P., Wayne, P.M., Stevenson, L.W., Wood, M.J., Forman, D., Davis,
R.B., Phillips, R.S, 2011. Tai Chi exercise in patients with chronic heart failure. Arch
Intern Med, 171, 750–757.
Yeh, G.Y., Wood, M.J., Lorell, H.B., Stevenson, L.W., Eisenberg, D.M., Wayne, P.M.,
Goldberger, A.L., Davis, R.B., Phillips, R.S, 2004. Effects of Tai Chi mind-body
movement therapy on functional status and exercise capacity in patients with chronic
heart failure: a randomized controlled trial. Am J Med, 117, 541–548.
26
Tables
Table 1 Demographic, Taiji-related, and psychometric group characteristics of the 49 subjects under study who
completed the Trier Social Stress Test (TSST)
Group characteristics
Taiji group
(n = 26)
Control group
(n = 23) p
Age 1 (years) 35.77 ±1.61 35.74 ±1.31 .99
Gender (male / female) 8 / 18 9 /14 .56
Menstrual cycle phase at TSST examination day
(female subjects in luteal phase / in follicular phase) 9 / 9 4 / 10 .29
Body mass index (kg/m2) 23.47 ±.67 23.51 ±.79 .97
Education
(with / without high school degree i.e. Swiss “Matura”) 20 / 6 17 / 6 1.00
Occupational status
(full or part time workers / students) 24 / 2 23 / 0 .49
Smoking
(non smokers / light smokers2)
21 / 5 18 / 5 1.00
Sportive activity – during the intervention
(hrs/week) 2.40 ±.37 2.98 ±.52 .37
Previous experience with mind-body practices
(months of regular practice; pre intervention) 15.62 ±6.03 29.13 ±10.68 .26
Depressive affect (ADS-K score) 10.88 ±1.35 11.04 ±1.45 .94
Perceived stress (PSS score) 17.46 ±1.03 17.91 ±1.22 .78
Taiji classes attended (incl. %-value) 20.65 ±.59 (86%) - -
Taiji practice at home (number of sessions/week) 1.69 ±.33 - -
1 Continuous data are expressed as mean ± S.E.M.
2 smoking less than 5 cigarettes per day
27
Table 2 Psychological reactivity in response to the Trier Social Stress Test (TSST) 1
Variables
Taiji Group (n = 26) Control Group (n = 23)
p 3
Pre TSST Post TSST
Stress
change 2
Pre TSST Post TSST
Stress
change b
Self-reported
stressfulness
(VAS) 4
1.22 .19 3.47 .45 2.25 .39 1.01 .20 4.91 .48 3.90 .41 .006
Calmness
(MDMS) 5
16.47 .46 12.78 .62 -3.69 .54 16.25 .49 10.64 .66 -5.61 .57 .019
Mood
(MDMS) 5
17.30 .44 14.23 .74 -3.07 .64 17.19 .47 12.39 .79 -4.80 .68 .070
1 All data are expressed as mean ± S.E.M.
2 Stress change = post TSST value minus pre TSST value
3 p-values refer to repeated measures ANCOVAs with prior mind-body practice as a covariate
4 VAS = visual analogue scale ranging from 0 = ‘not stressful at all’ to 10 = ‘extremely stressful’
5 MDMS = multidimensional mood scale ranging from 5 to 20 with higher scores indicating a higher degree of
calmness and a more positive mood
28
Figure Legends
Legend to Figure 1
Flow diagram for the progress through the phases of the randomized trial (based on the
consolidated standards of reporting trials [CONSORT] recommendations).
Legend to Figure 2
Values are means S.E.M. We calculated ANCOVAs with repeated measures of
physiological stress parameters as dependent variables and group (Taiji vs. Control) as the
independent variable. We controlled for physiological baseline level, age, gender, menstrual
cycle phase and prior experience with mind-body practices as covariates. The Taiji group
showed attenuated cortisol stress reactivity (p = .028; Figure 2a), α-amylase levels (p = .049;
Figure 2b), as well as lower heart rate stress responses (p = .028; Figure 2c). Significance
levels are: ° = p < .1; * = p < .05; ** = p < .01).
29
Figure 1 Documentation of the subject inclusion process
30
Figure 2a
31
Figure 2b
32
Figure 2c